Source Digest — Caplan, The Myth of the Rational Voter

Status (April 2026): Complete standard digest. Two thematic clusters: (1) rational irrationality and systematic voter biases; (2) implications for democracy-as-capture that differ from Friedberg's.

Source identification

Thematic cluster 1: rational irrationality and systematic biases

Core claims

• The standard public-choice story of "rational ignorance" (voters don't know because learning is costly and a single vote decides nothing) is incomplete. It predicts random error, not systematic bias.
• Voters are not merely ignorant; they are systematically biased. Because voting's decisiveness is effectively zero, the cost of holding a comforting-but-false belief is near zero, so voters indulge "rational irrationality" — the consumption of preferred beliefs.
• Caplan identifies four systematic biases that diverge from expert (professional-economist) consensus:
  1. Anti-market bias — underestimating market coordination benefits.
  2. Anti-foreign bias — underestimating gains from trade and migration.
  3. Make-work bias — equating prosperity with employment rather than production.
  4. Pessimistic bias — perceiving economic conditions as worse than they are.
• The result: democracies choose worse policies than median-voter models predict, because the median voter's preferences are systematically distorted relative to factual economic analysis.

Representative excerpt (from Cato PA 594)

"In a democracy, the primary function of government is not to serve the median voter's interests. It is to serve the median voter's beliefs about her interests. When those beliefs are systematically mistaken, democracies systematically choose suboptimal policies — not by accident, but by design."

Thematic cluster 2: democracy-as-capture, but of a different kind

Core claims

• Caplan's democracy-as-capture is categorically different from Friedberg's. Friedberg's argument is that beneficiary blocs vote for more government. Caplan's is that voters across the spectrum hold biased beliefs about how economies work, independent of their transfer status.
• Under Caplan's account, beneficiary-bloc capture and rational-irrationality capture can both hold simultaneously. They reinforce each other: biased voters have low information about the long-run fiscal consequences of transfers, so transfers expand further than median-voter preferences (correctly informed) would predict.
• Caplan proposes modest remedies: more decision-making by expert-delegated bodies (monetary policy, antitrust, specific regulatory agencies) and less by direct electoral referendum. He is explicit that this is in tension with democratic legitimacy norms.

Research context

Interpretive notes

• Caplan's work is the most rigorous libertarian statement of democratic-capture-by-electorate-ignorance, and it is genuinely different from the beneficiary-bloc capture story Friedberg offers. The exchange should treat these as two distinct capture mechanisms, both partially true.
• The Round 1 constructive analysis already notes that the project should treat democracy as contested terrain, not guaranteed substrate. Caplan gives that caution empirical spine: the project cannot assume "more democratic process" automatically corrects "bad policy" if the electorate is systematically biased on specific questions.
• The project's response to Caplan is naturally to distinguish domains. In domains where expert consensus is strong and stable (monetary policy, disease surveillance), delegation has a defensible track record. In domains where "expert consensus" is itself contested (redistribution, immigration, industrial policy), delegation is more about whose experts are empowered than about rationality correction.

Project 2028 mapping

• Exchange: Government Overreach, Ownership as Transition, and the Ratchet Problem. Directly relevant to Question 3 of Round 2 priorities (democracy as capture). Adds an electorate-belief mechanism to Friedberg's beneficiary-bloc mechanism.
• Problem Map: Domain 3 (Information ecosystems), Domain 15 (Democratic process). Caplan's rational-irrationality framework names a class of §15 failures that are not downstream of §3 manipulation alone — voters with very low decision stakes can rationally indulge biases, which constrains how much §3 reform can buy on its own.
• Principles: Tests Principle 14 (truth and evidence must be protected as public goods). If voter beliefs are systematically biased, §14's emphasis on truth-infrastructure matters more — not less — for democratic capture-resistance.
• Round 2 use: Use to force the exchange to separate "beneficiary-bloc capture" from "belief-systematic capture." The project's answers to each will differ.

Cross-references